Tracking devices can track people and objects in real time. These devices typically ascertain information relating to their physical locations based on communications with a variety of different wireless locationing systems (e.g., the Global Positioning System (GPS), cellular network systems (e.g., GSM), and wireless local area networks (e.g., a system of Wi-Fi access points). No single approach, however, provides continuous tracking information under all circumstances. For example, the GPS tracking requires a tracking device to have an unobstructed view of at least four GPS satellites at the same time, making GPS tracking in urban and indoor environments problematic. A variety of locationing techniques have been developed for tracking in indoor environments, including vision-based localization, wireless based localization (e.g., Received Signal Strength Indicator (RSSI) triangulation and fingerprinting techniques), and acoustic background fingerprinting. Each of these techniques, however, requires certain infrastructure support (e.g., wireless access points at known locations) and/or prior knowledge of the target environment (e.g., predetermined fingerprint maps) and therefore is associated with its own set of problems and limitations. Dead reckoning locationing based on motion sensor measurements also may be used, but the locationing accuracy of this approach is limited.
Tracking devices that incorporate multiple locationing mechanisms have been proposed to provide localization functionality across heterogeneous environments, ranging from environments equipped with localization equipment (e.g., satellites, cellular towers, and wireless access points), to environments without any localization equipment. However, incorporating a variety of different locationing components into a tracking device poses significant integration difficulties and challenges optimizing weight, size, cost, and battery life for a given application. In addition, localization techniques used in areas without any infrastructure support are notoriously inaccurate and typically require a person to physically traverse a storage facility until the person is close enough to the target to distinguish the signals (e.g., RFID signals) emitted from multiple co-located tracking devices. Thus, there still remains a need to address the lack of sufficient infrastructure to support continuous tracking across different environments.